数据分析

部署环境

(base) $ conda config --add channels conda-forge
(base) $ conda config --set channel_priority strict

(base) $ conda create -y -n pydata-book python=3.10
(base) $ conda activate pydata-book    //进入pydata-book

安装依赖

题外：尽量全部使用conda安装，不成功再用pip

(pydata-book) $ conda install -y pandas jupyter matplotlib

conda install lxml beautifulsoup4 html5lib openpyxl requests sqlalchemy seaborn scipy statsmodels patsy scikit-learn pyarrow pytables numba

IPython

(pydata-book) ipython

In [1]: a = 5

In [2]: a
Out[2]: 5

In [3]: import numpy as np

In [4]: data = [np.random.standard_normal() for i in range(7)]

In [5]: data
Out[5]: 
[-0.20470765948471295,
 0.47894333805754824,
 -0.5194387150567381,
 -0.55573030434749,
 1.9657805725027142,
 1.3934058329729904,
 0.09290787674371767]

Jupyter Notebook

(pydata-book) Jupyter Notebook

• 重命名：单击页面顶部的笔记本标题并键入新标题，完成后按 Enter

数据结构

• 元组：元素不能随意更改

  • 字符串转元祖

    tup = tuple('string')
    

  • 变量拆包：

    In [34]: values = 1, 2, 3, 4, 5
    
    In [35]: a, b, *rest = values
    
    In [36]: a
    Out[36]: 1
    
    In [37]: b
    Out[37]: 2
    
    In [38]: rest
    Out[38]: [3, 4, 5]
    

• 列表

  • append
  • insert
  • pop
  • sort
  • 切片

• 字典

  • del
  • pop
  • update

• 集合

python基础

• 匿名 (Lambda) 函数

  • equiv_anon = lambda x: x * 2
    

  • In [201]: def apply_to_list(some_list, f):
       .....:     return [f(x) for x in some_list]
    
    In [202]: ints = [4, 0, 1, 5, 6]
    
    In [203]: apply_to_list(ints, lambda x: x * 2)
    Out[203]: [8, 0, 2, 10, 12]
    

• 打开文件

  • In [233]: path = "examples/segismundo.txt"
    
    In [234]: f = open(path, encoding="utf-8")
    

  • In [237]: f.close()
    

pytorch

发展

• Theano
  • TensorFlow
  • TensorFloweager
• Torch7
  • pytorch+THNN
  • pytorch+Caffe2
• Caffe
  • Caffe2
  • Pytorch1.0